Publish by:

Keynote Speakers

Precision Motion Control and Morphing Mechanisms

Richard W. Longman
Columbia University
New York, NY 10027
USA

Abstract

        Engineering advances are often made at the boundary between two fields. This paper creates synergy between the design of mechanisms used in manufacturing equipment and the design of control systems. Particular attention is given to cam follower systems that are normally designed under the assumption that there is a control system that produces a constant rotation speed. But the variation in the cam resistance to rotation over a revolution makes the speed vary, and hence the design does not precisely perform as intended. Iterative learning control (ILC) and repetitive control (RC) are intelligent control techniques that can address this problem by learning to eliminate periodic errors. But they can go much further. If the cam is imperfectly machined, they can fix the resulting behavior and make it behave as if it were perfectly machined – morphing the cam from the actual hardware to the intended design. If at some time one would like to operate the equipment at a higher speed, one can morph the mechanism behavior to be that of a different cam, one designed for the higher speed. And this can be accomplished in software without hardware modifications.

×
  Topic : Precision Motion Control and Morphing Mechanisms
Abstract
Richard W. Longman    
Richard W. Longman

AIAA Fellow
AAS Fellow
Professor of Mechanical Engineering
Columbia University in New York City, USA
     
   


Optimizing Control – Intelligent Technologies for Automotive Engineering


Hans Georg Bock
University of Heidelberg
Germany

Abstract
×

        Optimizing Control is a relatively new control approach to ensure an intelligent operation of complex – nonlinear - dynamic technological processes in all kinds of application areas. Its applicability includes areas as diverse as automotive engineering or the operation of processing units in chemical and biochemical production plants.


Its basic methodology is :
  • to employ on a sufficiently detailed, possibly first principles, mathematical model of the dynamic process,
  • and to perform an online dynamic optimization of the process controls and operating parameters in order to achieve an economically optimal operation,
  • while meeting various constraints on, e.g., emissions, wear, forces and torques, safety requirements or operating limits of the equipment.

        For long time dynamic processes subject to perturbations, such as a heavy duty truck moving from one destination to the next, the optimization is typically performed over a finite prediction horizon and repeated with sufficiently high sampling rate taking into account the latest state and parameter estimates, which defines a nonlinear, optimal and constrained feedback control (NMPC). One computational challenge is the sufficiently fast solution of the dynamic programming problem to be real-time feasible that has recently been met by the introduction of so-called “multi-level real-time iterations”.

        Another feature of the new approach is the incorporation of “mixed-integer” control variables, which can assume only discrete values. Examples are gears in vehicle dynamics, open or closed valves, or any logical decision such as switching on or off additional devices coupled to the dynamic process. While classical algorithms for nonlinear mixed-integer optimization are far too slow to be real-time feasible for complex fast processes, new methods based on an “outer convexification” are fast enough to solve even the time optimal control for a complete lap of a race car in real-time. The paper will present applications for race cars and heavy duty trucks that have led to a Daimler patent.

        These methods have recently also been demonstrated to be feasible for chemical and biochemical production units, both in continuous operation and for the start-up and shut-down phase.

  Topic : Optimizing Control - Intelligent Technologies For Automotive Engineering
Abstract
   
Hans Georg Bock

Co-Chairman, strategic Committee for Mathematical Modeling, Simulation and Optimization (KoMSO) at the Federal Ministry of Education and Science

Interdisciplinary Center for Scientific Computing IWR
University of Heidelberg, Germany
Hans Georg Bock
   


Topic : Wearable Pneumatic Control Components and Its Application to Rehabilitation Field.

Shujiro Dohta
Okayama University of Science, Japan
Professor of Department of Intelligent Mechanical Engineering
Vice-president, Okayama University of Science, Japan

Abstract
×
  Topic : Wearable Pneumatic Control Components and Its Application to Rehabilitation Field.
Abstract
Pau-Choo Chung    
Shujiro Dohta

Okayama University of Science, Japan
Professor of Department of Intelligent Mechanical Engineering
Vice-president, Okayama University of Science, Japan
     
   







 

Powered by OpenConf®
Copyright ©2002-2012 Zakon Group LLC